cpython/Lib/test/test_functools.py

586 lines
20 KiB
Python

import functools
import sys
import unittest
from test import support
from weakref import proxy
import pickle
from random import choice
@staticmethod
def PythonPartial(func, *args, **keywords):
'Pure Python approximation of partial()'
def newfunc(*fargs, **fkeywords):
newkeywords = keywords.copy()
newkeywords.update(fkeywords)
return func(*(args + fargs), **newkeywords)
newfunc.func = func
newfunc.args = args
newfunc.keywords = keywords
return newfunc
def capture(*args, **kw):
"""capture all positional and keyword arguments"""
return args, kw
def signature(part):
""" return the signature of a partial object """
return (part.func, part.args, part.keywords, part.__dict__)
class TestPartial(unittest.TestCase):
thetype = functools.partial
def test_basic_examples(self):
p = self.thetype(capture, 1, 2, a=10, b=20)
self.assertEqual(p(3, 4, b=30, c=40),
((1, 2, 3, 4), dict(a=10, b=30, c=40)))
p = self.thetype(map, lambda x: x*10)
self.assertEqual(list(p([1,2,3,4])), [10, 20, 30, 40])
def test_attributes(self):
p = self.thetype(capture, 1, 2, a=10, b=20)
# attributes should be readable
self.assertEqual(p.func, capture)
self.assertEqual(p.args, (1, 2))
self.assertEqual(p.keywords, dict(a=10, b=20))
# attributes should not be writable
if not isinstance(self.thetype, type):
return
self.assertRaises(AttributeError, setattr, p, 'func', map)
self.assertRaises(AttributeError, setattr, p, 'args', (1, 2))
self.assertRaises(AttributeError, setattr, p, 'keywords', dict(a=1, b=2))
p = self.thetype(hex)
try:
del p.__dict__
except TypeError:
pass
else:
self.fail('partial object allowed __dict__ to be deleted')
def test_argument_checking(self):
self.assertRaises(TypeError, self.thetype) # need at least a func arg
try:
self.thetype(2)()
except TypeError:
pass
else:
self.fail('First arg not checked for callability')
def test_protection_of_callers_dict_argument(self):
# a caller's dictionary should not be altered by partial
def func(a=10, b=20):
return a
d = {'a':3}
p = self.thetype(func, a=5)
self.assertEqual(p(**d), 3)
self.assertEqual(d, {'a':3})
p(b=7)
self.assertEqual(d, {'a':3})
def test_arg_combinations(self):
# exercise special code paths for zero args in either partial
# object or the caller
p = self.thetype(capture)
self.assertEqual(p(), ((), {}))
self.assertEqual(p(1,2), ((1,2), {}))
p = self.thetype(capture, 1, 2)
self.assertEqual(p(), ((1,2), {}))
self.assertEqual(p(3,4), ((1,2,3,4), {}))
def test_kw_combinations(self):
# exercise special code paths for no keyword args in
# either the partial object or the caller
p = self.thetype(capture)
self.assertEqual(p(), ((), {}))
self.assertEqual(p(a=1), ((), {'a':1}))
p = self.thetype(capture, a=1)
self.assertEqual(p(), ((), {'a':1}))
self.assertEqual(p(b=2), ((), {'a':1, 'b':2}))
# keyword args in the call override those in the partial object
self.assertEqual(p(a=3, b=2), ((), {'a':3, 'b':2}))
def test_positional(self):
# make sure positional arguments are captured correctly
for args in [(), (0,), (0,1), (0,1,2), (0,1,2,3)]:
p = self.thetype(capture, *args)
expected = args + ('x',)
got, empty = p('x')
self.assertTrue(expected == got and empty == {})
def test_keyword(self):
# make sure keyword arguments are captured correctly
for a in ['a', 0, None, 3.5]:
p = self.thetype(capture, a=a)
expected = {'a':a,'x':None}
empty, got = p(x=None)
self.assertTrue(expected == got and empty == ())
def test_no_side_effects(self):
# make sure there are no side effects that affect subsequent calls
p = self.thetype(capture, 0, a=1)
args1, kw1 = p(1, b=2)
self.assertTrue(args1 == (0,1) and kw1 == {'a':1,'b':2})
args2, kw2 = p()
self.assertTrue(args2 == (0,) and kw2 == {'a':1})
def test_error_propagation(self):
def f(x, y):
x / y
self.assertRaises(ZeroDivisionError, self.thetype(f, 1, 0))
self.assertRaises(ZeroDivisionError, self.thetype(f, 1), 0)
self.assertRaises(ZeroDivisionError, self.thetype(f), 1, 0)
self.assertRaises(ZeroDivisionError, self.thetype(f, y=0), 1)
def test_weakref(self):
f = self.thetype(int, base=16)
p = proxy(f)
self.assertEqual(f.func, p.func)
f = None
self.assertRaises(ReferenceError, getattr, p, 'func')
def test_with_bound_and_unbound_methods(self):
data = list(map(str, range(10)))
join = self.thetype(str.join, '')
self.assertEqual(join(data), '0123456789')
join = self.thetype(''.join)
self.assertEqual(join(data), '0123456789')
def test_pickle(self):
f = self.thetype(signature, 'asdf', bar=True)
f.add_something_to__dict__ = True
f_copy = pickle.loads(pickle.dumps(f))
self.assertEqual(signature(f), signature(f_copy))
class PartialSubclass(functools.partial):
pass
class TestPartialSubclass(TestPartial):
thetype = PartialSubclass
class TestPythonPartial(TestPartial):
thetype = PythonPartial
# the python version isn't picklable
def test_pickle(self): pass
class TestUpdateWrapper(unittest.TestCase):
def check_wrapper(self, wrapper, wrapped,
assigned=functools.WRAPPER_ASSIGNMENTS,
updated=functools.WRAPPER_UPDATES):
# Check attributes were assigned
for name in assigned:
self.assertTrue(getattr(wrapper, name) is getattr(wrapped, name))
# Check attributes were updated
for name in updated:
wrapper_attr = getattr(wrapper, name)
wrapped_attr = getattr(wrapped, name)
for key in wrapped_attr:
self.assertTrue(wrapped_attr[key] is wrapper_attr[key])
def _default_update(self):
def f(a:'This is a new annotation'):
"""This is a test"""
pass
f.attr = 'This is also a test'
def wrapper(b:'This is the prior annotation'):
pass
functools.update_wrapper(wrapper, f)
return wrapper, f
def test_default_update(self):
wrapper, f = self._default_update()
self.check_wrapper(wrapper, f)
self.assertIs(wrapper.__wrapped__, f)
self.assertEqual(wrapper.__name__, 'f')
self.assertEqual(wrapper.attr, 'This is also a test')
self.assertEqual(wrapper.__annotations__['a'], 'This is a new annotation')
self.assertNotIn('b', wrapper.__annotations__)
@unittest.skipIf(sys.flags.optimize >= 2,
"Docstrings are omitted with -O2 and above")
def test_default_update_doc(self):
wrapper, f = self._default_update()
self.assertEqual(wrapper.__doc__, 'This is a test')
def test_no_update(self):
def f():
"""This is a test"""
pass
f.attr = 'This is also a test'
def wrapper():
pass
functools.update_wrapper(wrapper, f, (), ())
self.check_wrapper(wrapper, f, (), ())
self.assertEqual(wrapper.__name__, 'wrapper')
self.assertEqual(wrapper.__doc__, None)
self.assertEqual(wrapper.__annotations__, {})
self.assertFalse(hasattr(wrapper, 'attr'))
def test_selective_update(self):
def f():
pass
f.attr = 'This is a different test'
f.dict_attr = dict(a=1, b=2, c=3)
def wrapper():
pass
wrapper.dict_attr = {}
assign = ('attr',)
update = ('dict_attr',)
functools.update_wrapper(wrapper, f, assign, update)
self.check_wrapper(wrapper, f, assign, update)
self.assertEqual(wrapper.__name__, 'wrapper')
self.assertEqual(wrapper.__doc__, None)
self.assertEqual(wrapper.attr, 'This is a different test')
self.assertEqual(wrapper.dict_attr, f.dict_attr)
def test_missing_attributes(self):
def f():
pass
def wrapper():
pass
wrapper.dict_attr = {}
assign = ('attr',)
update = ('dict_attr',)
# Missing attributes on wrapped object are ignored
functools.update_wrapper(wrapper, f, assign, update)
self.assertNotIn('attr', wrapper.__dict__)
self.assertEqual(wrapper.dict_attr, {})
# Wrapper must have expected attributes for updating
del wrapper.dict_attr
with self.assertRaises(AttributeError):
functools.update_wrapper(wrapper, f, assign, update)
wrapper.dict_attr = 1
with self.assertRaises(AttributeError):
functools.update_wrapper(wrapper, f, assign, update)
@unittest.skipIf(sys.flags.optimize >= 2,
"Docstrings are omitted with -O2 and above")
def test_builtin_update(self):
# Test for bug #1576241
def wrapper():
pass
functools.update_wrapper(wrapper, max)
self.assertEqual(wrapper.__name__, 'max')
self.assertTrue(wrapper.__doc__.startswith('max('))
self.assertEqual(wrapper.__annotations__, {})
class TestWraps(TestUpdateWrapper):
def _default_update(self):
def f():
"""This is a test"""
pass
f.attr = 'This is also a test'
@functools.wraps(f)
def wrapper():
pass
self.check_wrapper(wrapper, f)
return wrapper
def test_default_update(self):
wrapper = self._default_update()
self.assertEqual(wrapper.__name__, 'f')
self.assertEqual(wrapper.attr, 'This is also a test')
@unittest.skipIf(not sys.flags.optimize <= 1,
"Docstrings are omitted with -O2 and above")
def test_default_update_doc(self):
wrapper = self._default_update()
self.assertEqual(wrapper.__doc__, 'This is a test')
def test_no_update(self):
def f():
"""This is a test"""
pass
f.attr = 'This is also a test'
@functools.wraps(f, (), ())
def wrapper():
pass
self.check_wrapper(wrapper, f, (), ())
self.assertEqual(wrapper.__name__, 'wrapper')
self.assertEqual(wrapper.__doc__, None)
self.assertFalse(hasattr(wrapper, 'attr'))
def test_selective_update(self):
def f():
pass
f.attr = 'This is a different test'
f.dict_attr = dict(a=1, b=2, c=3)
def add_dict_attr(f):
f.dict_attr = {}
return f
assign = ('attr',)
update = ('dict_attr',)
@functools.wraps(f, assign, update)
@add_dict_attr
def wrapper():
pass
self.check_wrapper(wrapper, f, assign, update)
self.assertEqual(wrapper.__name__, 'wrapper')
self.assertEqual(wrapper.__doc__, None)
self.assertEqual(wrapper.attr, 'This is a different test')
self.assertEqual(wrapper.dict_attr, f.dict_attr)
class TestReduce(unittest.TestCase):
func = functools.reduce
def test_reduce(self):
class Squares:
def __init__(self, max):
self.max = max
self.sofar = []
def __len__(self):
return len(self.sofar)
def __getitem__(self, i):
if not 0 <= i < self.max: raise IndexError
n = len(self.sofar)
while n <= i:
self.sofar.append(n*n)
n += 1
return self.sofar[i]
def add(x, y):
return x + y
self.assertEqual(self.func(add, ['a', 'b', 'c'], ''), 'abc')
self.assertEqual(
self.func(add, [['a', 'c'], [], ['d', 'w']], []),
['a','c','d','w']
)
self.assertEqual(self.func(lambda x, y: x*y, range(2,8), 1), 5040)
self.assertEqual(
self.func(lambda x, y: x*y, range(2,21), 1),
2432902008176640000
)
self.assertEqual(self.func(add, Squares(10)), 285)
self.assertEqual(self.func(add, Squares(10), 0), 285)
self.assertEqual(self.func(add, Squares(0), 0), 0)
self.assertRaises(TypeError, self.func)
self.assertRaises(TypeError, self.func, 42, 42)
self.assertRaises(TypeError, self.func, 42, 42, 42)
self.assertEqual(self.func(42, "1"), "1") # func is never called with one item
self.assertEqual(self.func(42, "", "1"), "1") # func is never called with one item
self.assertRaises(TypeError, self.func, 42, (42, 42))
self.assertRaises(TypeError, self.func, add, []) # arg 2 must not be empty sequence with no initial value
self.assertRaises(TypeError, self.func, add, "")
self.assertRaises(TypeError, self.func, add, ())
self.assertRaises(TypeError, self.func, add, object())
class TestFailingIter:
def __iter__(self):
raise RuntimeError
self.assertRaises(RuntimeError, self.func, add, TestFailingIter())
self.assertEqual(self.func(add, [], None), None)
self.assertEqual(self.func(add, [], 42), 42)
class BadSeq:
def __getitem__(self, index):
raise ValueError
self.assertRaises(ValueError, self.func, 42, BadSeq())
# Test reduce()'s use of iterators.
def test_iterator_usage(self):
class SequenceClass:
def __init__(self, n):
self.n = n
def __getitem__(self, i):
if 0 <= i < self.n:
return i
else:
raise IndexError
from operator import add
self.assertEqual(self.func(add, SequenceClass(5)), 10)
self.assertEqual(self.func(add, SequenceClass(5), 42), 52)
self.assertRaises(TypeError, self.func, add, SequenceClass(0))
self.assertEqual(self.func(add, SequenceClass(0), 42), 42)
self.assertEqual(self.func(add, SequenceClass(1)), 0)
self.assertEqual(self.func(add, SequenceClass(1), 42), 42)
d = {"one": 1, "two": 2, "three": 3}
self.assertEqual(self.func(add, d), "".join(d.keys()))
class TestCmpToKey(unittest.TestCase):
def test_cmp_to_key(self):
def mycmp(x, y):
return y - x
self.assertEqual(sorted(range(5), key=functools.cmp_to_key(mycmp)),
[4, 3, 2, 1, 0])
def test_hash(self):
def mycmp(x, y):
return y - x
key = functools.cmp_to_key(mycmp)
k = key(10)
self.assertRaises(TypeError, hash(k))
class TestTotalOrdering(unittest.TestCase):
def test_total_ordering_lt(self):
@functools.total_ordering
class A:
def __init__(self, value):
self.value = value
def __lt__(self, other):
return self.value < other.value
self.assert_(A(1) < A(2))
self.assert_(A(2) > A(1))
self.assert_(A(1) <= A(2))
self.assert_(A(2) >= A(1))
self.assert_(A(2) <= A(2))
self.assert_(A(2) >= A(2))
def test_total_ordering_le(self):
@functools.total_ordering
class A:
def __init__(self, value):
self.value = value
def __le__(self, other):
return self.value <= other.value
self.assert_(A(1) < A(2))
self.assert_(A(2) > A(1))
self.assert_(A(1) <= A(2))
self.assert_(A(2) >= A(1))
self.assert_(A(2) <= A(2))
self.assert_(A(2) >= A(2))
def test_total_ordering_gt(self):
@functools.total_ordering
class A:
def __init__(self, value):
self.value = value
def __gt__(self, other):
return self.value > other.value
self.assert_(A(1) < A(2))
self.assert_(A(2) > A(1))
self.assert_(A(1) <= A(2))
self.assert_(A(2) >= A(1))
self.assert_(A(2) <= A(2))
self.assert_(A(2) >= A(2))
def test_total_ordering_ge(self):
@functools.total_ordering
class A:
def __init__(self, value):
self.value = value
def __ge__(self, other):
return self.value >= other.value
self.assert_(A(1) < A(2))
self.assert_(A(2) > A(1))
self.assert_(A(1) <= A(2))
self.assert_(A(2) >= A(1))
self.assert_(A(2) <= A(2))
self.assert_(A(2) >= A(2))
def test_total_ordering_no_overwrite(self):
# new methods should not overwrite existing
@functools.total_ordering
class A(int):
pass
self.assert_(A(1) < A(2))
self.assert_(A(2) > A(1))
self.assert_(A(1) <= A(2))
self.assert_(A(2) >= A(1))
self.assert_(A(2) <= A(2))
self.assert_(A(2) >= A(2))
def test_no_operations_defined(self):
with self.assertRaises(ValueError):
@functools.total_ordering
class A:
pass
class TestLRU(unittest.TestCase):
def test_lru(self):
def orig(x, y):
return 3*x+y
f = functools.lru_cache(maxsize=20)(orig)
domain = range(5)
for i in range(1000):
x, y = choice(domain), choice(domain)
actual = f(x, y)
expected = orig(x, y)
self.assertEquals(actual, expected)
self.assert_(f.hits > f.misses)
self.assertEquals(f.hits + f.misses, 1000)
f.clear() # test clearing
self.assertEqual(f.hits, 0)
self.assertEqual(f.misses, 0)
f(x, y)
self.assertEqual(f.hits, 0)
self.assertEqual(f.misses, 1)
# Test bypassing the cache
self.assertIs(f.__wrapped__, orig)
f.__wrapped__(x, y)
self.assertEqual(f.hits, 0)
self.assertEqual(f.misses, 1)
# test size zero (which means "never-cache")
@functools.lru_cache(0)
def f():
nonlocal f_cnt
f_cnt += 1
return 20
f_cnt = 0
for i in range(5):
self.assertEqual(f(), 20)
self.assertEqual(f_cnt, 5)
# test size one
@functools.lru_cache(1)
def f():
nonlocal f_cnt
f_cnt += 1
return 20
f_cnt = 0
for i in range(5):
self.assertEqual(f(), 20)
self.assertEqual(f_cnt, 1)
# test size two
@functools.lru_cache(2)
def f(x):
nonlocal f_cnt
f_cnt += 1
return x*10
f_cnt = 0
for x in 7, 9, 7, 9, 7, 9, 8, 8, 8, 9, 9, 9, 8, 8, 8, 7:
# * * * *
self.assertEqual(f(x), x*10)
self.assertEqual(f_cnt, 4)
def test_main(verbose=None):
test_classes = (
TestPartial,
TestPartialSubclass,
TestPythonPartial,
TestUpdateWrapper,
TestTotalOrdering,
TestWraps,
TestReduce,
TestLRU,
)
support.run_unittest(*test_classes)
# verify reference counting
if verbose and hasattr(sys, "gettotalrefcount"):
import gc
counts = [None] * 5
for i in range(len(counts)):
support.run_unittest(*test_classes)
gc.collect()
counts[i] = sys.gettotalrefcount()
print(counts)
if __name__ == '__main__':
test_main(verbose=True)